|Publication number||US5399513 A|
|Application number||US 07/979,562|
|Publication date||Mar 21, 1995|
|Filing date||Nov 17, 1992|
|Priority date||Jun 27, 1989|
|Also published as||DE69028159D1, DE69028159T2, EP0405293A1, EP0405293B1|
|Publication number||07979562, 979562, US 5399513 A, US 5399513A, US-A-5399513, US5399513 A, US5399513A|
|Inventors||Tian-I Liou, Chih-Sieh Teng|
|Original Assignee||National Semiconductor Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (15), Non-Patent Citations (22), Referenced by (26), Classifications (12), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This is a continuation of application Ser. No. 07/762,644, filed on Sep. 19, 1991, now abandoned, which is a FWC of parent application Ser. No. 371,483, filed Jun. 27, 1989, now abandoned.
The invention herein relates to the production of CMOS devices. More particularly it relates to a method for production of CMOS devices in which certain masks can be eliminated.
A significant limiting factor in current n-channel MOSFETs is hot-electron-induced degradation, because of high channel electric fields and electron impact ionization rates in the micron and sub-micron n-channel devices. In hot electron injection, electrons are injected into the gate oxide by the high electric field created by the short channel region adjacent to the drain. As a result, the threshold voltage of the device is altered. In addition, overlap between the gate electrode and the source and drain results in parasitic capacitance between the diffusion region and the gate. This phenomenon, called Miller capacitance, decreases device speed.
"Lightly doped drain" (LDD) and "double diffused drain" (DDD) structures have been developed to minimize this degradation. Both of these structures are based on the theory that if there is a lightly doped n-surface layer adjacent to the channel, the peak electric field will be reduced and shifted from under the gate so that hot electron-induced degradation will be lessened. However, both pose other types of problems. Production of LDD structures requires the use of an additional mask to prevent LDD phosphorus from implanting into p-channel regions. In both the LDD and DDD structures one must construct a highly doped n+ region to produce low contact resistance at the n-channel source and drain for high-speed circuits.
The production of CMOS (complementary metal-oxide semiconductor) devices requires somewhere in the area of eleven masks to produce the sequential layers of oxides, active areas and contacts which form the device. Each step in the device production generally involves a separate mask, except where "blanket" implant and oxidations can be effected. Each mask which must be used adds to the cost and time required to produce the device. The requirement of added mask steps to reduce hot electron effects has provided greater incentive to find other steps for which masks may be eliminated.
One effort to decrease the number of mask steps involved the use of differential oxidation to eliminate the p+ implant mask. (A. Hui, et al, "An Oxide Masked P+ Source/Drain Implant for VLSI CMOS", IEDM 1982 pp 698.) An n+ implant is performed with a masking step to cover p+ regions with photoresist. Oxide is grown after the photoresist is stripped. Because of the difference in dopant concentrations between the p- and the n- channel source/drain regions, a thicker oxide is grown over the n-channel regions. A p+ implant can be performed which will pass through the thin oxide over the p-channel sources and drains but will be blocked by the thicker oxide over the n-channel sources and drains. This process, however, is not compatible with a self-aligned silicide (salicide) process.
Salicide processes are desirable for their ability to produce low sheet resistance at contacts to polysilicon, n+ and p+ regions for MOS devices in the micron and submicron ranges. Salicides allow for increased device speeds where scaling down of contact dimensions would otherwise result in higher contact resistance, causing slower devices. Steps must be taken, however, to provide a clean silicon surface for silicide formation, and to prevent shorts between the gate and the source and/or drain. Such steps are not provided in the prior proposed differential oxide process.
It would therefore be advantageous to be able to combine a procedure which allows for the elimination of a mask step and provides protection against hot-electron effects with a salicide process to provide low resistance contacts and increased device speed. This is a preferred embodiment of the present invention.
The present process comprises the use of a differential oxidation of the source/drain regions to permit elimination of the p+ implant mask normally required for formation of p-channel device in a CMOS process. A DDD procedure provides protection against hot-electron effects. A second oxide spacer is included to allow formation of salicide at the contacts to provide low sheet resistance.
After completion of a standard CMOS self-aligned gate process through formation of the oxide spacers on the edges of the gate structures, sequential implants of a low energy, high dosage first n-type dopant and a low energy, medium dosage second n-type dopant are performed. A differential thermal oxide is grown, with an accelerated growth rate over the n-channel regions. A p-type dopant is implanted with medium energy and high dosage into p-channel regions. The differential thermal oxide is etched back to silicon, and a set of oxide spacers is formed at the edges of the gate structures. A salicide is formed in the active areas to provide low contact resistance.
The single Figure of the drawing is a flow diagram illustrating the sequence of steps in the production process of the present invention.
The process of this invention can be best understood by reference to the Figure. The process for fabrication of self-aligned gate MOS devices is well known to those skilled in the art. Therefore, the process steps prior to the description of the present invention will not be detailed, but are merely included to complete the process flow. The proposed process incorporates several novel steps which markedly improve the ease of production while yielding a CMOS device which has operating characteristics equivalent to or better than devices made by conventional processes.
The present process starts with a p-type substrate 2 as illustrated in the Figure. The preferred embodiment uses a twin well process. The n-well 6 is formed by an n- implant into the area exposed by selective removal of a composite layer of thin SiO2 and Si3 N4. A thick oxide is grown in the n- area, after which the Si3 N4 is stripped. A p-well 7 is formed by implant into the area where the Si3 N4 was etched away. The n- area is masked against the p- implant by the thicker oxide. All oxides are then stripped and the two wells are driven in.
A first alternate embodiment is the formation of only an n-well in a p-type substrate. The "p-well" in this case would simply be the area which was not exposed during n-implant. A second alternate embodiment would involve starting with a n-type substrate. A p-well would be formed using a similar masking step through which a p-type dopant would be implanted. Both alternate embodiments are widely used in the industry, however, excessive doping effects involved in single-well processes leads to the preference of a twin well process for the present invention.
The active regions 4 are subsequently defined by use of a composite-Si3 N4 mask 3, and then are followed by p-field implant 8 using a field implant blocking mask.
Device isolation is achieved by field oxidation 10 of non-active areas through a mask of oxidation-inhibiting silicon nitride, or a composite thereof. Following removal of the silicon nitride, a gate oxide 12 is grown in the active areas, followed by a Vt channel implant 14. Polysilicon 16 is deposited and doped, and is etched as defined by the polysilicon mask 18 to form the gate structures 19. All steps to this point are conventional for CMOS self-aligned gate technology. Slight variations in this procedure may be anticipated and will not impact the scope of the invention.
A low temperature oxide is deposited followed by anisotropic etching to form oxide spacers 22 adjacent to the edges of the gate structures 19. (As an alternative, a thermal oxide may be grown followed by anisotropic etching to form oxide spacers.) Photoresist 20 is spun on the substrate and patterned to expose the n-channel source/drain areas. A dual implant of n-type dopant species is performed to create a double-diffused drain structure in the source/drain regions 24 of the n-channel devices. The first implant 21 is a high dosage, low energy arsenic implant for low contact resistance. The second implant 23 is medium dosage, low energy phosphorous, which diffuses along a shallow area below the surface, forming a lightly doped drain adjacent to the area below gate structure 19. For the purposes of this process, "low" dosage will be defined as doses of <1013 atoms/cm2, "medium" dosage as 1013 -1015 atoms/cm2 and "high" dosage as >1015 atoms/cm2.
A thermal oxidation is performed to grow a differential oxide. The oxide 26 on the n-channel source/drain regions 24 is about 1000Å due to the oxidation-rate enhancement of the n-type dopants. The oxide 28 on the p-channel source/drain 30 grows to only about 400Å. This thinner oxide allows the p+ implant 32 to pass through it into the underlying substrate. In areas where the oxide is thicker, such as the field oxide 10 and oxide 26 over the n-channel source/drain regions 24, the implant is blocked sufficiently to prevent the p-type dopant from entering the substrate.
Thereafter, the structure is subjected to rapid thermal annealing (RTA) to drive in the implants and activate the dopants. Furnace annealing at 900° C. for approximately 80 minutes provides the same degree of dopant activation and counterdoping, however, RTA is preferred in order to minimize diffusion of the junctions.
The differential oxide 26 and 28 is etched back to the underlying substrate. Care must be taken in this step in selection of plasma etch gases and power to avoid damaging the silicon in the p-channel source/drain regions. The etchback will also attack the oxide spacers 22, so it is desirable to provide another step to form spacers for the subsequent salicide steps.
A second deposition of low temperature oxide is performed, followed by anisotropic etching to form oxide spacers 34 adjacent to the edges of the gate structures 19. (The alternative of thermal oxidation to form spacers is not available here, as it is undesirable to subject the source/drain regions to unnecessary high temperature processing to avoid diffusion of the implants.) These spacers serve to prevent shorting from the gate to the source and/or drain as a result of the formation of the salicide.
A thin film of titanium 36 is deposited and the source/drain regions are heat treated to form a salicide 38 of TiSi2. The titanium 36 does not react where it is deposited on oxide, specifically over .field oxide 10 and oxide spacers 34. The unreacted titanium is subsequently removed. The salicide 38 produces low sheet resistance in all contacts.
A low temperature dielectric film 40 is deposited and patterned with photoresist through which contact openings 42 are etched. Interconnect 44 is achieved using conventional procedures of metal deposition and patterning.
By this technique, the need to use LDD and p+ masks is eliminated. The inclusion of a second set of spacers allows the formation of a salicide to improve device speed.
Specific examples derived from experimental results will illustrate the process of this invention. The double diffused source/drain is created by a sequential implant of arsenic and phosphorous. The arsenic implant had a dose of 4×1015 atoms/cm2 and an energy of 50 KeV. It is contemplated that doses in the range of 2×1015 to 8×1015 atoms/cm2 will be suitable for use, with the preferred ranges being 3×1015 to 6×1015 atoms/cm2. The phosphorous implant had a dose of 2×1014 atoms/cm2 and an energy of 50 KeV. Generally the phosphorus implant will be conducted with doses of 9×1013 to 4×1014 atoms/cm2 and energies of 40 to 60 KeV. The latter implant provides protection against hot-electron effects by forming a shallow, lightly doped junction adjacent to the device gate. The arsenic implant follows conventional processes, providing a low contact resistance in the n-channel sources and drains.
The differential oxidation was accomplished in an oxidation furnace at 900° C. for 40 minutes in an atmosphere of pure oxygen.
The p+ implant energy must be high enough to allow the dopant to pass through the 400Å differential oxide, but not so high as to enter the regions protected by the 1000Å differential oxide. A BF2 implant of 3×1015 atoms/cm2 and 70 KeV was used. A suitable range for this implant is 2×1015 to 5×1015 atoms/cm2 and 60 to 80 KeV. The device characteristics of a CMOS device made by the process of this invention are given in the Table below, derived from tests of 100/l μm MOSFET's.
TABLE______________________________________ n-channel p-channel______________________________________Vt, V 0.76 -0.87M-factora, V0.5 0.53 -0.425 × 5 Ib, mA 27.4 16.10BVdss, V 13.4 -12.1Isx (max)c, μA 49.5 0.10Betad (4V), μA/V2 71 20.6 Sheet Resistance, Junction ohm/square Depth, μm______________________________________n+ 3.2 0.17P+ 3.9 0.25polysilicon, n+ 6.6 --n- 2400 1.7______________________________________ Notes: a. Body effect b. Saturation current where VD = VG = 5 V c. Peak substrate current d. "BetaFactor" = μCox, where μ = mobility and Cox = gate oxide capacitance
It will be evident that there are additional embodiments which are not illustrated above but which are clearly within the scope and spirit of the present invention. The above description and drawings are therefore intended to be exemplary only and the scope of the invention is to be limited solely by the appended claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4314857 *||Nov 8, 1979||Feb 9, 1982||Mitel Corporation||Method of making integrated CMOS and CTD by selective implantation|
|US4356623 *||Sep 15, 1980||Nov 2, 1982||Texas Instruments Incorporated||Fabrication of submicron semiconductor devices|
|US4358890 *||Aug 18, 1981||Nov 16, 1982||Ibm Corporation||Process for making a dual implanted drain extension for bucket brigade device tetrode structure|
|US4470852 *||Sep 3, 1982||Sep 11, 1984||Ncr Corporation||Method of making CMOS device and contacts therein by enhanced oxidation of selectively implanted regions|
|US4474624 *||Jul 12, 1982||Oct 2, 1984||Intel Corporation||Process for forming self-aligned complementary source/drain regions for MOS transistors|
|US4480375 *||Dec 9, 1982||Nov 6, 1984||International Business Machines Corporation||Simple process for making complementary transistors|
|US4753898 *||Jul 9, 1987||Jun 28, 1988||Motorola, Inc.||LDD CMOS process|
|US4771014 *||Sep 18, 1987||Sep 13, 1988||Sgs-Thomson Microelectronics, Inc.||Process for manufacturing LDD CMOS devices|
|US4775642 *||Feb 2, 1987||Oct 4, 1988||Motorola, Inc.||Modified source/drain implants in a double-poly non-volatile memory process|
|US4949136 *||Jun 9, 1988||Aug 14, 1990||University Of Connecticut||Submicron lightly doped field effect transistors|
|US4963504 *||Nov 24, 1989||Oct 16, 1990||Xerox Corporation||Method for fabricating double implanted LDD transistor self-aligned with gate|
|US5091763 *||Dec 19, 1990||Feb 25, 1992||Intel Corporation||Self-aligned overlap MOSFET and method of fabrication|
|US5219784 *||Apr 2, 1990||Jun 15, 1993||National Semiconductor Corporation||Spacer formation in a bicmos device|
|EP0164449A2 *||Dec 24, 1984||Dec 18, 1985||Hitachi, Ltd.||Process for producing a semiconductor integrated circuit device including a MISFET|
|EP0216053A2 *||Jul 17, 1986||Apr 1, 1987||Motorola, Inc.||Removable sidewall spaces for lightly doped drain formation using one mask level|
|1||*||A. C. Hui et al, An Oxide Masked P Source/Drain Implant for VLSI CMOS, IEDM 82, pp. 698 701.|
|2||A. C. Hui et al, An Oxide Masked P+ Source/Drain Implant for VLSI CMOS, IEDM '82, pp. 698-701.|
|3||*||Chen, M L., et al., A High Performance Submicron CMOS Process with Self Aligned Chan Stop and Punch Through Implants (Twin Tub V), International Electron Devices Meeting Technical Digest, Dec. 7 10, 1986, pp. 256 259 (Los Angeles).|
|4||Chen, M-L., et al., "A High Performance Submicron CMOS Process with Self-Aligned Chan-Stop and Punch-Through Implants (Twin-Tub V)," International Electron Devices Meeting Technical Digest, Dec. 7-10, 1986, pp. 256-259 (Los Angeles).|
|5||*||Fair et al., Modeling Physical Limitations on Junction Scaling of CMOS, IEEE Trans. on Electron Devices, ED 31, No. 9, Sep. 1984, pp. 1180 1185.|
|6||Fair et al., Modeling Physical Limitations on Junction Scaling of CMOS, IEEE Trans. on Electron Devices, ED-31, No. 9, Sep. 1984, pp. 1180-1185.|
|7||*||H. Mikoshiba et al., Comparison of Drain Structures in n Channel MOSFETS, IEEE Trans. on Electron Dev., ED 33, No. 1, Jan. 1986, pp. 140 144.|
|8||H. Mikoshiba et al., Comparison of Drain Structures in n-Channel MOSFETS, IEEE Trans. on Electron Dev., ED-33, No. 1, Jan. 1986, pp. 140-144.|
|9||*||J. Agraz Guerena et al., Twin Tub III A Third Generation CMOS Technology, IEDM 84, pp. 63 66.|
|10||J. Agraz-Guerena et al., Twin-Tub III -A Third Generation CMOS Technology, IEDM '84, pp. 63-66.|
|11||*||L. C. Parrillo et al., Twin Tub CMOS A Technology for VLSI Circuits, IEDM 80, pp. 752 755.|
|12||*||L. C. Parrillo et al., Twin Tub CMOS II An Advanced VLSI Technology, IEDM 82, pp. 706 709.|
|13||L. C. Parrillo et al., Twin-Tub CMOS --A Technology for VLSI Circuits, IEDM '80, pp. 752-755.|
|14||L. C. Parrillo et al., Twin-Tub CMOS II --An Advanced VLSI Technology, IEDM '82, pp. 706-709.|
|15||*||Lai, Fang shi J., et al., A Highly Latchup Immune 1 m CMOS Technology Fabricated with 1 Me V Ion Implantation and Self Aligned TiSi 2 , IEEE Transactions on Electron Devices, vol. ED 33, No. 9, Sep. 1986, pp. 1308 1320 (New York).|
|16||Lai, Fang-shi J., et al., "A Highly Latchup-Immune 1-μm CMOS Technology Fabricated with 1-Me V Ion Implantation and Self-Aligned TiSi2," IEEE Transactions on Electron Devices, vol. ED-33, No. 9, Sep. 1986, pp. 1308-1320 (New York).|
|17||*||R. R. Doering et al., A High Performance 1 m CMOS Process for VLSI Applications, IEEE 1985 Custom Integrated Circuits Conference, pp. 196 198.|
|18||R. R. Doering et al., A High Performance 1 μm CMOS Process for VLSI Applications, IEEE 1985 Custom Integrated Circuits Conference, pp. 196-198.|
|19||*||Scott, et al., Considerations for Scaled CMOS Source/Drains, Reprint IEDM 81, pp. 538 541.|
|20||Scott, et al., Considerations for Scaled CMOS Source/Drains, Reprint -IEDM '81, pp. 538-541.|
|21||*||Wolf et al., Silicon Processing for the VLSI Era, vol. 1: Process Technology, Lattice Press, 1986, pp. 397 399.|
|22||Wolf et al., Silicon Processing for the VLSI Era, vol. 1: Process Technology, Lattice Press, 1986, pp. 397-399.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5585299 *||Mar 19, 1996||Dec 17, 1996||United Microelectronics Corporation||Process for fabricating a semiconductor electrostatic discharge (ESD) protective device|
|US5747372 *||Aug 5, 1996||May 5, 1998||Lg Semicon Co., Ltd.||Semiconductor device and method for fabricating same|
|US5849616 *||Dec 19, 1996||Dec 15, 1998||Mitsubishi Denki Kabushiki Kaisha||Method of manufacturing a semiconductor device|
|US5963784 *||May 9, 1997||Oct 5, 1999||Vlsi Technology, Inc.||Methods of determining parameters of a semiconductor device and the width of an insulative spacer of a semiconductor device|
|US6368986 *||Aug 31, 2000||Apr 9, 2002||Micron Technology, Inc.||Use of selective ozone TEOS oxide to create variable thickness layers and spacers|
|US6503851||Aug 3, 2001||Jan 7, 2003||Micron Technology, Inc.||Use of linear injectors to deposit uniform selective ozone TEOS oxide film by pulsing reactants on and off|
|US6602807||Nov 18, 2002||Aug 5, 2003||Micron Technology, Inc.||Use of linear injectors to deposit uniform selective ozone TEOS oxide film by pulsing reactants on and off|
|US6617230||Sep 18, 2001||Sep 9, 2003||Micron Technology, Inc.||Use of selective ozone teos oxide to create variable thickness layers and spacers|
|US6638824||Jan 11, 2002||Oct 28, 2003||Fairchild Semiconductor Corporation||Metal gate double diffusion MOSFET with improved switching speed and reduced gate tunnel leakage|
|US6844225||Jan 15, 2003||Jan 18, 2005||International Business Machines Corporation||Self-aligned mask formed utilizing differential oxidation rates of materials|
|US6888176||Jun 26, 2003||May 3, 2005||T-Ram, Inc.||Thyrister semiconductor device|
|US7192893||Aug 5, 2003||Mar 20, 2007||Micron Technology Inc.||Use of linear injectors to deposit uniform selective ozone TEOS oxide film by pulsing reactants on and off|
|US7214979||Aug 25, 2004||May 8, 2007||Micron Technology, Inc.||Selectively deposited silicon oxide layers on a silicon substrate|
|US7279367||Dec 7, 2004||Oct 9, 2007||T-Ram Semiconductor, Inc.||Method of manufacturing a thyristor semiconductor device|
|US7288827||Oct 20, 2004||Oct 30, 2007||International Business Machines Corporation||Self-aligned mask formed utilizing differential oxidation rates of materials|
|US7804107||Oct 3, 2007||Sep 28, 2010||T-Ram Semiconductor, Inc.||Thyristor semiconductor device and method of manufacture|
|US7968886 *||Jan 29, 2009||Jun 28, 2011||Semiconductor Energy Laboratory Co., Ltd.||Semiconductor integrated circuit and method of fabricating same|
|US8288219 *||Mar 20, 2008||Oct 16, 2012||Hynix Semiconductor, Inc.||Method of forming a non-volatile memory cell using off-set spacers|
|US8946003||Feb 20, 2007||Feb 3, 2015||SK Hynix Inc.||Method of forming transistors with ultra-short gate feature|
|US20020098633 *||Jan 31, 2002||Jul 25, 2002||William Budge||Use of selective ozone TEOS oxide to create variable thickness layers and spacers|
|US20040029402 *||Aug 5, 2003||Feb 12, 2004||William Budge|
|US20040137670 *||Jan 15, 2003||Jul 15, 2004||International Business Machines Corporation||Self-aligned mask formed utilizing differential oxidation rates of materials|
|US20050035418 *||Aug 25, 2004||Feb 17, 2005||Micron Technology, Inc.||Selectively deposited silicon oxide layers on a silicon substrate|
|US20050079726 *||Oct 20, 2004||Apr 14, 2005||International Business Machines Corporation||Self-aligned mask formed utilizing differential oxidation rates of materials|
|US20080166844 *||Mar 20, 2008||Jul 10, 2008||Hynix Semiconductor, Inc.||Method of Forming a Non-volatile Memory Cell Using Off-set Spacers|
|US20090134462 *||Jan 29, 2009||May 28, 2009||Semiconductor Energy Laboratory Co., Ltd.||Semiconductor integrated circuit and method of fabricating same|
|U.S. Classification||438/227, 438/305, 257/E21.636, 438/231, 257/E27.067|
|International Classification||H01L21/265, H01L27/092, H01L29/78, H01L21/8238|
|Cooperative Classification||H01L21/823835, H01L27/0928|
|Sep 21, 1998||FPAY||Fee payment|
Year of fee payment: 4
|Sep 20, 2002||FPAY||Fee payment|
Year of fee payment: 8
|Oct 9, 2002||REMI||Maintenance fee reminder mailed|
|Sep 21, 2006||FPAY||Fee payment|
Year of fee payment: 12